How do you know if you are at a local optimum?
If incremental changes yield minimal improvement and radically different approaches have not been explored, you may be on a local peak.
Mental Models
Local vs. Global Optima
A local optimum is the best solution within a narrow neighborhood; a global optimum is the best solution across the entire possibility space.
This model prevents teams from getting trapped in good-enough solutions by reminding them that the landscape of possibilities may have much better peaks elsewhere.
Core question
Are we optimizing within our current approach, or have we explored whether a fundamentally different approach is better?
Example
A team squeezes 2% more conversion from their existing checkout flow. Meanwhile, a competitor redesigns checkout from scratch and achieves 40% higher conversion. The incumbent was stuck on a local peak.
How to apply
- 1.Recognize when incremental improvements are yielding diminishing returns.
- 2.Periodically explore radically different approaches, even if current performance is acceptable.
- 3.Use small experiments to test distant alternatives before committing.
- 4.Accept short-term performance dips when climbing toward a higher peak.
Pitfalls
- ·Constantly chasing global optima and never finishing anything.
- ·Abandoning a strong local position without evidence of something meaningfully better.
- ·Underestimating transition costs between peaks.
Questions
When is a local optimum good enough?
When the cost of finding and transitioning to a global optimum exceeds the expected benefit, or when time pressure requires shipping now.
Related
- Diminishing Returns
Each additional unit of effort eventually yields less incremental value.
- First Principles Thinking
Reason from fundamentals instead of analogy alone.
- Premature Optimization
Optimizing before understanding the problem wastes effort on the wrong things.